How would you implement a multi-layer cache strategy to efficiently fetch token prices for 1000 users accessing a widget simultaneously?

Imagine you have a lemonade stand and lots of people want lemonade at the same time. You keep some lemonade in a small cooler nearby for quick access, more in a bigger cooler a bit further away, and even more in a big fridge at home. This way, you can serve everyone quickly without running out.